This invention relates to electronic learning aids, teaching machines and electronic games. More particularly, this invention relates to electronic learning aids, teaching machines or games having means for producing synthesized speech implemented in a minature semiconductor chip.
In the prior art various electronic teaching devices and games are known. For example, a small electronic learning aid for teaching arithmetic to children using randomly selected problems is disclosed in U.S. Pat. No. 3,584,398. Further, teaching machines are known which rely on traditional movie film or video tape techniques for presenting both audio and visual information to a student and would include means for posing questions to the student and receiving and correcting answers from the student. A proposal for such an automatic teaching device is found in the Paul K. Weimer article in "IRE Transactions on Education" of June 1958. It should be evident, however, that a teaching machine employing movie projectors or video tape machines is bulky, heavy and fairly expensive to manufacture. Furthermore, it is desirable to at least partially randomize the questions posed by the learning aid; this function is, of course, difficult to implement using conventional audio or video tape machines or movie projectors.
The prior art also suggests various techniques for synthesizing human speech from digital data. For instance, some of the techniques used are briefly described in "Voice Signals: Bit by Bit" at pages 28-34 of the Oct. 1973 issue of IEEE Spectrum. An important technique for synthesizing human speech, and the technique used by the speech synthesizer chip described herein, is called linear predictive coding. For a detailed discussion of this technique, see "Speech Analysis and Synthesis by Linear Prediction of the Speech Wave" by B. S. Atal and Suzanne L. Hanauer which appears at pp. 637-50 of Volume 50, No. 2 (part 2) (1971) of the Journal of the Acoustical Society of America.
In U.S. patent application Ser. No. 807,461 filed June 17, 1977, abandoned in favor of continuation application Ser. No. 905,328 filed May 12, 1978, now U.S. Pat. No. 4,209,844 issued June 24, 1980, a lattice filter capable of being implemented on a single semiconductor chip as described. The speech synthesis chip described herein makes use of the lattice filter described in the aforementioned U.S. Pat. No. 4,209,844.
It is one object of this invention that the learning aid or game be equipped to audibly ask questions of the user thereof.
It is another object of this invention that the teaching machine receive an answer to a posed question from the opertor and to inform the operator whether or not the inputted answer is correct.
It is still yet another object of this invention that the questions posed be randomly selectable.
It is yet another object of this invention that speech synthesis circuits be implemented on a miniature semiconductor chip so that the entire talking learning aid be simply constructable as a portable device.
The foregoing objects are achieved as is now described. The questions to be posed by the machine are stored as digital codes in a memory device. This memory is preferably of the non-volatile type so that the questions posed are not erased when power is disconnected from the apparatus. A speech synthesizer circuit is connected to the output of the memory for selectively converting the digital signals stored therein to speech signals from which audible speech is generated. Several types of speech synthesis circuits are known. In the disclosed embodiment, the speech synthesizer is implemented using linear predictive coding and integrated on a single semiconductor chip. A speaker or earphone and an amplifier (if needed) are provided to convert the output from the speech synthesizer to audible sounds. A keyboard and display, both of which preferably are capable of accommodating alphanumeric characters, are preferably provided. The display and keyboard are preferably coupled to the speech synthesis circuit and memory via a controller circuit. In the embodiment disclosed, the controller circuit is an appropriately programmed microprocessor device. The controller circuit controls the memory to read out the digital signals corresponding to a question to be posed, the question preferably being randomly selected from a plurality of questions stored therein. The question posed is converted to audible signals by means of the synthesizer circuit in combination with the speaker or earphone. The memory also preferably stores data indicative of the correct answer to the question posed, which data is supplied to the controller circuit. When the operator answers the questions posed by inputting his or her answer at the keyboard, the controller compares the inputted answer with the answer stored in the memory and the operator is informed of the results of this comparison. The operator may be so informed either visually via the display or audibly via the speech synthesis circuit and speaker or earphone, to inform the operator "very good", for example, if the operator gave an incorrect answer. The question posed may, of course, be either a rather complex, lengthy question or alternatively, as in the case of the disclosed embodiment, may be as simple as speaking a word and awaiting a correct spelling thereof. Of course, the shorter the questions posed the greater the number of questions storable in a memory of given capacity. The learning aid is preferably arranged to have several levels of difficulty. Thus the easiest level might have such words as "dog", "cat", "time", and the like while the next level might have words such as "mother", "flower", and the like and so forth. Of course, the particular words selected for any given library are a design choice. The controller circuit preferably controls from which difficulty level the posed question is to be randomly selected. The particular difficulty level used is selected based on instructions inputted at the keyboard or by other means. After the operator gives a correct answer, e.g. the correct spelling of the word "spoken" then the learning aid goes on to preferably select another random word. When an incorrect answer is given, the controller circuit preferably causes the word to be posed again after the operator is informed that the answer is incorrect and if the operator continues to give an incorrect answer, the controller circuit provides via the display or the speech synthesis circuit the correct answer and then goes on to randomly select another word or question to be posed. The learning aid is installed in an easily portable case. The size of the speech synthesis chip is on the order of 45,000 square mils using conventional MOS design rules and a conventional P-MOS processing technique. Of course, using C-MOS processing would tend to increase the size of the chip somewhat.
In the embodiment disclosed, the learning aid is preferably equipped with other modes of operation which are described in detail.